Invisimail doesn't consistently process email addresses

Track this issue

You are looking at the html source, correct? If everything works right, there shouldn't be any visual difference on the page.

Using browser view source, I am seeing a plaintext email address not an encoded one.

I have many email addresses that are not being obfuscated or autolinked. I cannot determine a rational pattern as to why some are processed and others are not, although it is repeatable (non-random). Some emails are consistently not processed, while others (on the same page) are. I'm running the current CVS version of the 4.7 branch.

Here's a strange test case. I have Textile installed.

1jim@fbz.com

Jim Foo
Foo, Bar & Baz, P.C.
335 N. South Way
Blankville, Utopia 55555
(800) 555-1212
(800) 555-5555(fax)
2jim@fbz.com
Admitted 1996 (MN); 2002 (AZ)

3jim@fbz.com

If I put Invisimail first in the filter ordering, only the 2jim address is processed. If I put Invisimail second (behind Textile), then only the 1jim1 and 3jim addresses are processed.

If I run Invisimail with only the line break converter (Invisimail -10, Line Break Converter 10), only the 2jim address is processed.

I'm going to try the regex from urlfilter in this module, or perhaps add an invisimail option to the urlfilter module.

The regex from urlfilter does indeed work. Perhaps email encoding could be a urlfilter option?

There are still cases where the email isn't processed. The crazy thing is, it's working for me locally under XAMPP, but on my server some emails are not processed unless I add a space before them when they are at the beginning of a line. Another strange thing is sometimes I can add a single letter and the email will be processed. Take that letter out and the email isn't processed. I'm at a loss as to how to debug this.

I also have some email addresses that are failing under my remote server and XAMPP. Something about the block of text they're in.

I updated the regex in the previous patch to take into account the fact that

tags may contain style information etc.

You may want to try to rearrange your filters so that "Encode email addresses" is processed first.

Go to administer > input formats, click on the Configure link of the input format that invisimail is being used with, and click on the "rearrange" tab.

Hope it helps.

I've tried it with all different orders (see my Post #6 above). Thanks.

Why does the regex care about

and

Sorry didn't know HTML was enabled in comments. First line should read: Why does the regex care about <p> and <li>?

4.7 is no longer supported. The 6.x-1.1 release has a new, more robust regex that should catch any reasonable case.